Public service agencies like hospitals, fire, rescue, and police departments are required to maintain a high level of service. These service standards often come as reliability constraints. For example, fire-related incidents require a 90% response rate in 4 minutes. We consider a case study of tackling out-of-hospital cardiac events using AED-enabled drones in Portland, OR. Environmental factors, majorly wind speed and direction, significantly impact drone performance. We formulate the drone location problem as a robust multi-period maximum coverage facility location problem. We model the coverage reliability constraint as a chance constraint on failure probabilities. For our context, multiple periods translate to periods with different wind speeds and distributions. The results show that extending to a multi-period formulation, rather than using average information in a single period, is particularly beneficial when either response time is short or uncertainty in failure probabilities is not accounted for. Accounting for uncertainty in decision-making improves coverage significantly while reducing variability, especially when response times are longer. Using multiple periods and accounting for uncertainty in failure probabilities boosts the simulated coverage values by 57%, on average.
Darshan Chauhan, Civil & Environmental Engineering, Portland State University
Darshan is a Ph.D. Candidate in Civil and Environmental Engineering and Graduate Research Assistant at Portland State University (PSU). His doctoral work with Dr. Avinash Unnikrishnan is on planning and real-time resource allocation in freight logistics systems using robust optimization and reinforcement learning. He is also interested in data analytics and its applications to transportation safety using econometric modeling. He has served as the Treasurer of STEP, PSU’s ITE student chapter.
Chauhan, Darshan, "PSU Student Research from the TRB 2022 Annual Meeting: Drone Facility Location Considering Coverage Reliability: Application to Emergency Medical Scenarios" (2022). TREC Friday Seminar Series. 218.